(155c) Olefin Plant Digitalization: Model-Based Coke Monitoring, Sensing of KPIs and Optimization for Steam Cracking Furnaces

The steam cracking furnace is at the heart of the ethylene production process, and its optimal operation is key to maximizing profitability. Accurate information on key process indicators, such as conversion, product yields, tube metal temperatures, and coking rates are essential for the optimal operation and control of furnaces. However, measuring these reliably in real time is often either impossible or problematic, requiring model based digital applications for reliable and timely information. As the furnace KPIs are affected by coking, it is also essential to have an accurate estimate of state of coking for accurate KPI prediction.

In addition, because feedstock availability changes with time, in order to maximize profitability it is necessary to periodically identify optimal operating conditions for the furnaces that nevertheless ensure operation within the plant constraints. Such optimization calculations should take into account the current state of coke build up in the furnaces to get correct results.

In this paper, we present three state-of-the-art Digital Applications for steam cracking furnaces to provide model-based coke build-up monitoring, real-time sensing of KPIs and optimization of operating conditions. Built on Siemens PSE’s gPROMS Digital Applications Platform, they use sophisticated numerical manipulations to combine rigorous first principles models of steam cracking furnaces with plant operating data. Experience with the deployment of these applications on full-scale industrial plants will be presented. In addition, concepts of more advanced optimization applications such as multiperiod mixed-integer optimization for production and decoke scheduling over a near-term time horizon, and integrated optimization of multiple olefin crackers, will be briefly discussed.